I had a friend loan me his 10-22, and I ended up purchasing the siggy. Granted the 10-22 is a wee bit faster, it's not $250 faster to me. I have had my siggy since august, and I love it. I find that the amount of distortion depends 100% on how I hold the camera in relation to an object. I can make the horizon look bent, or I can let it appear straight, I really like how it tends to distort the perspective making things close to the lens look huge and things further away look tiny.
While I don't agree with his overall review of the sigma, this link is a pretty good writeup on ultrawide lenses. My only concern is how long is the warranty on the siggies? I thought that it was only 1 year, but I have heard that it's 3 or 4.... regardless, I love the lens and can do a lot with it, here are some examples, you can check out my flickr account at www.flickr.com/photos/mtbtrials if you wanna try and suss out more of my sigma photos.
Canon 10-22 on the first three shots here, all three at 10mm...
Notice that this first one looks all out of wack because I took it from a low POV
Thanks fred, I just wanted to try and show an unbiased review of both lenses.
Unfortunatley my friend who owns the 10-22 somehow managed to damage the lens and I cannot do a side by side comparison.
I did test out the 10-20 before I made the purchase from my local camera shop (paid $500 for it from a *small* local shop), but I was happy with the lens when I walked out the store.
Also keep in mind that the sigma comes with a petal hood, and a nice pouch. I am not certain that the canon comes with them. Can anyone comment on this for me?
-Mark
Thanks for the tip for the 3 year warranty, I'm headed over there now.
I wanted to throw up two more images, not because they are great or anything, but I wanted to illustrate the sigma's ability to make the horizon appear straight (but on the first one, look at how the telegraph poles seem to converge)
Keep in mind that the first one was handheld while driving, and the second one was intended to be deleted, but I skipped over it. I was on my way home from detroit when I accidently ended up taking the wrong freeway (I was mad with my GPS) and made the most out of a photo op.
Bushwaker, I hope that my 10-20 images have helped, if you want me to take my lens outside and shoot on a tripod at given apetures I can do that either with my 300D or my dad's 400D. I think that the 400D also has mirror lock, and I downloaded the firmware for my 300D so I can lock the mirror.
I am pretty sure that the 10-20 is an all metal chassis. I know that the mount is, and it's bloody heavy!
Fred Lindsey wrote:
Yes but it is not lens distortion which is what we are comparing here - it happens to any lens of this focal range.
"ALL ultra wide angle lenses will produce image "distortion" if the horizon is not centered and level in the frame"
If we're still comparing the distortion in fisheye and ultrawide... well the distortion in the fisheye is by design... and lens distortion or not, the image taken with the ultrawide when it's not level with the horizon still end up looking unnatural and needing some sort of correction, unless you like your buildings looking skewed like that. Wasn't that your main argument against the fisheye? I guess the ultrawide has the same problem, only to a lesser degree.
Fred Lindsey wrote:
I don't think fisheyes can handle landscapes unless for creative effect or unless you're shooting at 0 degrees horizontal and vertical - this means no rule of thirds for the horizon.
Fred Lindsey wrote:
Yes it is 102-64 but to be honest it requires no PP and it has incredible IQ.
n0b0 wrote:
If we're still comparing the distortion in fisheye and ultrawide... well the distortion in the fisheye is by design... and lens distortion or not, the image taken with the ultrawide when it's not level with the horizon still end up looking unnatural and needing some sort of correction, unless you like your buildings looking skewed like that. Wasn't that your main argument against the fisheye? I guess the ultrawide has the same problem, only to a lesser degree.
Well, no, they're nothing like the same, and it's somewhat ridiculous to claim they are.
miccullen wrote:
Well, no, they're nothing like the same, and it's somewhat ridiculous to claim they are.
If you meant how the actual distortion look in the image then yes, they're not the same. I'm very much aware of that and it's insulting to think otherwise.
I think they fisheye shots look good and could think of some uses for them. In the end, though, unless you want the fisheye look, you will waste resolution and the width gained in defishing the images. The main reason I didn't buy the 10-17 fish is that I don't want to shoot fisheyes every time I shoot ultrawide, or even a majority of the time. The rest of the time 15mm will have to do.
danmitchell wrote:
The poster could be speaking of a couple of different types of distortion.
Barrel and pincushion distortion will vary depending on the design and construction of the lens. I suspect that this is the type of distortion you refer to when you write that yours is "roughly 0-1% at 10mm."
ALL ultra wide angle lenses will produce image "distortion" if the horizon is not centered and level in the frame - the inward/outward tilting walls, etc. that we often see in shots made with such lenses. This isn't a "flaw" in the lens - it is a feature! :-)
what about Sigma compatiblity problems in future EOS bodies?
Seems like everyone here has the Sigma.... what about the Canon 10-22 pictures anyone? and please don't correct the any distortions.
oh about fisheyes? I don't like them...makes me dizzy looking at it....--- no offense man I just don't like it.... what I want to capture is natural perspective.
There should be no problems with compatibility. All sigmas are in line with current cameras. Even if they do miraculously not work with future models sigma will undoubtedly re-chip it to work
Bushwaker, the first three shots I posted were 10-22 without any corrections. As I had stated, I only had the lens for a short while as it was my friends who lives across town. Unfortunatley he managed to break it and still hasn't gotten it fixed.
Two more with the canon 10-22
This is inside a vertical lathe shot at f13 at 10mm with the canon 10-22. The horizontal lines are fine, but look at how the vertical lines bow out more towards the center.
And here is one with the 10-20 (sigma) of the final product, my exif tells me that this was f8 at 11mm so I must have zoomed in a little versus step in.
and here is one other one with the 10-22
Please note that were the walls meet the ceiling were constructed with a bit of a curve, so that isn't from the lens.
I really hope that these are helping, I am not trying to sway you one way or another, I honestly feel like 10mm is the same on both lenses as far as distortion is concerned. The only advantage canon would have had for me woudl be the extra 2mm on the narrow end (and I hardly ever shoot my 10-20 tight) and the extra half stop when wide open which might be nice, but I usually crank the ISO up another step to make up more of a difference.
I would just run with whichever one your wallet is more comfortable with. Me? I don't plan on shooting crop sensors forever, so I got the 10-20 to bide my time until I can afford a 5D (or better) and a 16-35L (and a bunch of other L glass). But it's up to you, my 10-20 is going to have to get me thru at least another 5 years of shooting with crop sensors, if I ever decide to go FF. I just wanted that crazy wide angle to play with, and I am happier having that $250 difference in my pocket towards my new (insert what the other bit of kit you want here)
One other note about the canon 10-22 shot is I can't remember if it was shot with a tripod or not (the one of the naked bed and column), but the one of the finished machine was shot with my tripod.
One of the things I have yet to see mentioned about the Canon 10-22 is the CA. I had one of these way back with my 20D and could not *wait* to get rid of it. Horrible CA on every highlighted edge. Looked like purple feathers flocking the edges. Maybe it was just a bad one but several others have reported the same thing.
